The key to this is to get enough original image data and analyze the algorithm to get accurate results.
I learned that Oneview, an Israeli satellite data AI analysis company, has recently completed a multi-million dollar seed round of financing. This round of financing is jointly invested by MTK (MediaTek), Ondine Capital (active water capital), Israeli venture capital firm TPY Capital and a satellite industry-related enterprise.
Oneview was established in November 2018, mainly providing commercial satellite remote sensing image analysis services based on GAN and big data technology. Remote sensing is one of the main application scenarios of satellites. It refers to the observation of the earth through satellites to obtain information outside the scope of human vision. It has been used in military reconnaissance, surveying and mapping, and environmental resource monitoring in the past. However, with the current increase in satellite coverage frequency and accuracy, satellite remote sensing has gradually landed in the commercial field. For example, in the financial industry, some investors have begun to use remote sensing satellite images to monitor changes in the flow of cars in department stores, supermarkets, restaurants, etc., and make corresponding business decisions based on this. In the energy field, some customers can continuously monitor, collect and analyze data in a specified time zone to obtain information such as oil extraction, reserves, and production capacity estimates in the time zone. Compared with the traditional model, the method of obtaining results through remote sensing image analysis is more accurate, and can help the demand side to understand the required information in a shorter time.
From the perspective of business processes, Oneview ’s service method is to obtain the required original images from upstream remote sensing satellite manufacturers, and then perform data extraction, cleaning, analysis and interpretation based on the images to provide downstream customers The required data information and decision suggestions.
The key to doing this is to get enough original image data, and to analyze the algorithm to get accurate results. Image processing in the satellite field is different from other industries. Data analysts who want to obtain a specific result based on image analysis often need to work with several satellite manufacturers with different specifications to obtain data from different angles. And because of the small number of satellite images and the high price, data analysts must get enough data through various means in order to train their own algorithms. To solve these two problems, Oneview has established a global cooperation relationship with many upstream satellite manufacturers to ensure that the company can get more original images. Moreover, in its own algorithm training, Oneview matches the self-developed GAN algorithm in the existing satellite images to ensure the amount of training data through the generational confrontation network.
Moreover, the price of different satellite images will vary depending on factors such as resolution and shooting time. If the data analyst needs to obtain images with higher resolution or newer shooting time, it is best to establish contact with the supplier directly, but if the customer or some analysis needs require image clarity and shooting timeIf it is loose, you can get the image through the dealer. Regardless of the acquisition method, however, satellite data analysts like Oneview are required to produce analysis results that meet customer standards under existing conditions.
Oneview training example of machine learning algorithm based on self-developed algorithm
For customers, because there are not many suppliers available in the industry, the evaluation dimensions of various customers have not been unified-if customers have little experience in satellite remote sensing, they will choose through word of mouth supplier. If the customer has previous experience in using such services, they will select qualified suppliers according to their own purposes. In the selection, they may also test and compare the accuracy of the data of the new analyst based on the analysis results that have been previously available. Take the needs of existing commercial customers in the financial field as an example. The customer needs of this industry are mainly concentrated in statistics (such as financial institutions monitoring their own targets). Other types of needs include monitoring of people flow and logistics. After that, Oneview will expand the use of satellite data on the basis of existing services. For example, the same set of data can be used to analyze and derive new information from different angles. Such information may be provided to small and medium-sized customers by subscription and other means.
In the satellite remote sensing image analysis industry, other applications that can currently be seen appear in agricultural and ecological monitoring scenarios. As for whether it is possible for satellite image data analysts who are deeply cultivating agriculture and ecology to enter the commercial field, Oneview investors believe that the satellite image data analysis industry is still in the early stages and talents are scarce, so everyone is still at the stage of expanding the market together. And in terms of existing customer relationships, Oneview may have a first mover advantage.
For the team, the company ’s founder had previously worked in the Israeli military ’s science and technology department and had technical experience related to advanced satellite algorithms and ground stations. The company currently has about 10 employees, mainly technical R & D personnel. After this round of financing, Oneview plans to expand the team and further build a scaled software service platform.
The head company in the same field in the international market is the American remote sensing data service provider Orbital Insight, which has beenThe US $ 50 million in Series D financing was led by Sequoia Capital, Geodesic Capital, and Goldman Sachs Goldman Sachs (overseas). The company’s main customers are distributed in the fields of finance, energy, government agencies and non-profit organizations.